From glibc-2.2.x to glibc-2.3.x
highos at linuxfromscratch.org
Sat Oct 26 19:51:29 PDT 2002
On Sat, Oct 26, 2002 at 11:06:59PM -0300, Grant Leslie wrote:
> Just my 2 cents ( which since I'm Canadian is probably only worth about a buck
> twenty anyway *grin* )?
Yes, fear us Canadians... =)
> 1) One of the main reasons I want to use LFS, is to do things my own way, the
> "correct" way. And ( IMHO ) just because RedHat/Gentoo/Whoever did it one way,
> doesn't make it "correct"?
Neither does it make it incorrect.
I respect both Redhat and Gentoo. Not to mention any organization or
individual that takes time todo what they do. It is a long and boring
task and generally has little thrills. I perhaps may not use there
distrobutions, but that does not mean I should disrespect them, just
because we do not run the same system...
If anything we are at a disadvantage in this regard. Both Redhat and
Debian have developers that work on the GNU projects, by far much more
knowledgeable about theses pieces of software then any of us here.
That is not to say that instantly makes there choice correct. However
as I tried to point out, that does not mean it is incorrect either.
This is a pure technical transition problem. Bringing up this issue
wont solve it.
> 2) "If you want to hack your glibc 2.3 build to work around the problem,
> here"!! This is a "hack" too. But, also one that permanently affects your
> system. The change has been made to glibc for a reason, and they'll most
> likely not be reversing it in 2.3.2 and up, since it is, as they say
> "correct". Which leaves LFS in the exact same position the next time glibc
> gets upgraded, if the host system is LFS 4.0, or another patched distro, does
> it not? Say in LFS 6.0 building on RedHat 7.3, will we still need to patch
> glibc-2.4.2 to re-export those same symbols?
Greg is by far the most knowledgeable contributor on this subject, he
has spend far to much time doing research and testing on this issue.
And if I remenber correctly, what he proposed was actually the exact
same thing he first proposed when news of glibc 2.3 hit us.
So, yes.. reverting the changes in glibc 2.3 to what it was prior may
not be as "correct" as it should be. But it would make life far easier
for us in terms of maintaining the book and support issues.
If this were to take affect, it would be an optional patch anyways.
Some of us will be running glibc 2.3.x systems, in which case it will
not be required.
By far, all the other suggestions i've seen so far ressemble hacks that
send shivers up my back; more so then then what Greg has been
suggesting for the last month.
Uhh..anyways. Your remark in #1 kinda set off. Sorry if this comes off
a tad odd.. it's just that we are all part of the same grand community.
It's by far alot easier to leech of the ideas and work of others, then
it is to fight them over petty ideals :)
/me runs and hides now, wondering if perhaps he should have posted to
Jesse Tie-Ten-Quee ( highos at linuxfromscratch dot org )
Unsubscribe: send email to listar at linuxfromscratch.org
and put 'unsubscribe lfs-dev' in the subject header of the message
More information about the lfs-dev